Emag paramter retreival

Hey guys. Did anyone deal with emag parameter retrieval? I was doing MTM according to Smith's paper, using the same dimensions etc, and come up with pretty weird results. Namely, everything is kind of opposite. As an example I attach a refraction index plot. The real results from the paper shows up as imaginary in my simulation and vice versa. Anyone knows what I did wrong, suggestion or something?! Thanks...